Google DeepMind Releases Genie 4 and Project Genie Agent for AI Ultra Subscribers
Google DeepMind has rolled out Genie 4 alongside the Project Genie agent and a community gallery for Google AI Ultra subscribers, featuring real-time world generation and native audio.
By The Daily Circuit AI Desk (automated, AI-generated)
2 min read
AI-generated image (flux-1-schnell), not a photograph of the events described.
Key facts
- Google DeepMind rolled out Genie 4 for eligible Google AI Ultra subscribers.[1]
- Genie 4 introduces real-time natural language world editing and native adaptive audio.[1]
- Genie 3 updates include session extensions up to 3 minutes and 360-degree panorama grounding.[1]
- All generated video and audio in Project Genie includes SynthID watermarking.[1]
- Google AI Pro members can join a waitlist to access the new Project Genie experience.[1]
Google DeepMind has announced the rollout of Genie 4, its latest world model, alongside a new Project Genie experience tailored for eligible Google AI Ultra subscribers [1]. The update introduces real-time natural language environmental editing, native audio capabilities, a conversational creative agent, and a community sharing gallery [1].
New Capabilities in Genie 4 and Genie 3
Google DeepMind's Genie 4 model enables users to modify digital environments mid-session using natural language prompts [1]. Users can alter weather conditions, introduce new characters, or direct actions on the fly while navigating their created worlds [1]. Furthermore, the model incorporates real-time native audio that adapts dynamically as scenes evolve, ranging from ambient soundscapes to character dialogue [1].
Alongside Genie 4, updates to the Genie 3 model remain accessible within Project Genie's model picker [1]. Genie 3 now supports extended exploration sessions of up to three minutes per generation—an increase from the previous 60-second limit—as well as world chaining to link multiple environments together [1]. Additional Genie 3 updates include 360-degree panorama image grounding to seed spaces from panoramic photos and real-time promptable events during exploration [1].
Conversational Creation and Community Features
The newly introduced Project Genie agent allows users to build custom interactive universes through natural conversation rather than single text prompts [1]. The agent coordinates Google's multimodal models to establish visual styling and custom background scores [1]. Additionally, Gemini models can be brought into sessions to offer context-aware suggestion chips or trigger automated in-world events [1].
To foster collaboration, Google DeepMind has launched a community gallery where users can publish, explore, and remix custom worlds [1]. The launch features collaborations with pioneering creators, such as filmmaker Anand Gandhi's 'The Life of Ants' and creative technologist Will aka BLVCKL!GHT's 'Villa Giallo' [1].
Safety and Availability
Google DeepMind states that all generated video and audio within Project Genie incorporates SynthID watermarking, alongside safety screenings for model inputs and outputs [1]. Genie 3 and Genie 4 are positioned as frontier research models, with ongoing development addressing challenges such as camera latency, complex physical interactions, real-time audio precision, and long-term scene permanence [1]. Genie 4 is available immediately to eligible Google AI Ultra subscribers, while Google AI Pro members can join a waitlist [1].
Context
Project Genie was introduced earlier as an experimental research prototype by Google DeepMind to let users generate and explore interactive worlds directly from a web browser using advanced world models and Gemini models [1].
Why it matters
The release of Genie 4 and the Project Genie agent marks a significant advancement in interactive AI-generated media, combining real-time visual editing, native adaptive audio, and social sharing features for subscribers.
What could happen next
AI-generated analysis based on the sources below. Not a forecast or recommendation.
The integration of real-time world building and conversational agents could lower the technical barriers for digital content creation, paving the way for advanced interactive storytelling and virtual simulation tools.
Risks and open questions
Research challenges such as camera latency, intricate physical interactions, real-time audio precision, and long-term scene permanence remain active areas of development for frontier world models.
Sources
- 1.Google DeepMind (primary source): Genie 4 is now available, 18 November 2026
How this article was produced. This story was researched and written by an automated system using Google’s Gemini models, working only from the sources listed above. Automated checks verified that every citation, quote and figure matches those sources before publication. It was not written by a human reporter. Editorial policy · Report an error
- Gemini
- Genie 4
- Artificial Intelligence
- Google DeepMind
- World Models
Related stories
Google DeepMind Launches Gemini 4 Neon Model With 1 Million Token Context Window
Google DeepMind has introduced Gemini 4 Neon, a new model offering frontier-level intelligence, a 1 million-token context window, and advanced capabilities for software engineering, knowledge work, and science workflows.
2 min read
Nvidia Halts Advanced AI Chip Shipments to China Amid Leadership Changes
Nvidia has ceased shipments of its advanced H200 AI processors to China, according to statements made by interim CEO Colette Kress during an earnings call.
2 min read
Spotify Introduces Artist Profile Protection Tool to Tackle Deepfakes and Misattributed Music
Spotify is launching its Artist Profile Protection tool globally to help creators combat deepfakes, impersonation, and metadata errors on their streaming profiles.
2 min read
Nvidia Establishes National Security Board Committee Following Jensen Huang Resignation
Nvidia has formed a board-level committee to oversee national security policies and export compliance following the resignation of CEO Jensen Huang.
3 min read